Synaptic transmission and synaptic plasticity in the mouse striatum

(1)

From DEPARTMENT OF PHYSIOLOGY AND PHARMACOLOGY

Karolinska Institutet, Stockholm, Sweden

SYNAPTIC TRANSMISSION AND SYNAPTIC

PLASTICITY IN THE MOUSE STRIATUM

Sietske Schotanus

Stockholm 2008

(2)

Supervisor Faculty opponent

Docent Karima Chergui Docent Eric Hanse

Dept. of Physiology and Pharmacology Section for Physiology

Karolinska Institute University of Göteborg

Examination Board

Docent Peter Wallén

Dept.of Neuroscience

Karolinska Institute

Professor Bo Rydqvist

Dept. of Physiology and

Pharmacology

Karolinska Institute

Professor Per Andrén

Inst. for Pharmacological

Biosciences

Uppsala University

Cover illustration by Johanna M. Schotanus-van der Peijl Previously published papers were reproduced with permission from the publisher.

Published by Karolinska Institutet.

Printed by Universitetsservice US-AB.

(3)

"The greatest obstacle to discovery is not ignorance — it is the illusion of knowledge."

Daniel J. Boorstin

(4)

ABSTRACT

The striatum is the major input nucleus of the basal ganglia and can be subdivided into a dorsal part and a ventral part that is also named nucleus accumbens (NAc). The dorsal part is involved in motor control and habit learning whereas the ventral part is mostly associated with reward- motivated behaviors. The neurons that populate the striatum are for 95% GABAergic medium spiny neurons. Input into the striatum comes from cortex and thalamus and is mostly glutamatergic. This glutamatergic input is the essential drive behind excitatory synaptic transmission in the striatum. Apart from glutamatergic input, the striatum also receives dopaminergic input from the midbrain.

To measure excitatory synaptic transmission and synaptic plasticity we used field potential recordings which measures the activity of a population of neurons in the striatum evoked by stimulation of glutamatergic fibers. In order to study the involvement of specific neurotransmitters and receptors involved in synaptic transmission and its modulation, we applied pharmacological tools in the perfusion solution that modify glutamatergic, dopaminergic and GABAergic synaptic transmission. We also measured the levels of the neuromodulator dopamine which plays an important part in striatal synaptic transmission. In order to study long term potentiation (LTP), we applied a high frequency stimulation-protocol.

We found that glutamatergic synaptic transmission in the striatum is depressed by bath- application of N-methyl-D-aspartate (NMDA). We found that this depression is mediated by adenosine acting on A1-receptors. The NMDA-receptors that mediate this depression were shown to contain the NR1/NR2A-subnunits. These NMDA-receptors are most likely located in the striatum on medium spiny neurons. Furthermore, bath-applied NMDA also depresses evoked dopamine-release in striatum via NMDA-receptors that contain NR1/NR2A-subunits.

LTP in the NAc was shown to be independent of the Mg²⁺-block of NMDA receptors but rather depends on the level of NMDA-receptor activation. We also showed that LTP depends on dopamine D1- but not D2-receptors, is independent of GABA-receptor activation but requires the activation of group I mGluRs. Finally, we showed differences in neuronal excitability in the striatum and NAc between male and female mice in different stages of the estrous cycle. In addition, the excitability of striatal neurons of both male and female mice is modulated by acute administration of estrogen. Together, these results contribute to understanding the role of different neurotransmitters in the physiology of the striatum.

(5)

LIST OF PUBLICATIONS

This thesis is based on the following articles, referred to in the text by the roman numerals stated below:

I. Sietske M. Schotanus, Bertil B. Fredholm and Karima Chergui

NMDA depresses glutamatergic synaptic transmission in the striatum through the activation of adenosine A1 receptors: Evidence from knockout mice.

Neuropharmacology, 51 (2006) 272-282

II. Sietske M. Schotanus and Karima Chergui

NR2A containing N-methyl-D-aspartate receptors depress glutamatergic synaptic transmission and evoked dopamine-release in the mouse striatum.

Submitted to Journal of Neurochemistry

III. Sietske M. Schotanus and Karima Chergui

Long-term potentiation in the nucleus accumbens requires both NR2A- and NR2B-containing NMDA receptors

Submitted to European Journal of Neuroscience

IV. Sietske M. Schotanus and Karima Chergui

Dopamine D1 receptors and group I metabotropic glutamate receptors contribute to the induction of long-term potentiation in the nucleus accumbens.

Neuropharmacology, (2008) in press

(6)

LIST OF ABBREVIATIONS

AMPA amino-3-hydroxy-5-methyl-4-isoazole propionic acid CREB cAMP response element binding protein

(a)CSF (artificial) cerebrospinal fluid

DA Dopamine

DAT Dopamine transporter

DOPA Dihydroxyphenylalanine

DYN Dynorphin

EAAT Excitatory amino acid transporter

ENK Enkephalin

EP Entopeduncular nucleus

(f)EPSP/PS (field) Excitatory postsynaptic potential EPSC Excitatory postsynaptic current

ER Estrogen receptor

E2 Estrogen (17β-estradiol)

GABA Gamma (γ)-aminobutyric acid

GAD Glutamic acid decarboxylase

GAT GABA transporter

GP Globus pallidus

HFS High frequency stimulation

IPSP Inhibitory postsynaptic potential

KA Kainate

LTD Long term depression

LTP Long term potentiation

MAPK Mitogen-activated protein kinase mGluR Metabotropic glutamate receptor MSN Medium spiny (projection) neuron

NAc Nucleus accumbens

NMDA n-methyl-D-aspartate

RT Room temperature

SNc Substantia nigra pars compacta SNr Substantia nigra pars reticulata

SP Substance P

STN Subthalamic nucleus

VGLUT Vesicular glutamate transporter

VTA Ventral tegmental area

(9)

BACKGROUND

I. THE STRIATUM

The striatum is the major recipient of inputs into the basal ganglia. This sub-cortical nucleus participates in initiation, production and sequencing of movement and behavior and in the storage of information regarding these processes. A functional segregation can be made between the dorsal and ventral parts of the striatum, the ventral part being termed nucleus accumbens (NAc). The dorsal regions of the striatum have been closely linked to the production of task-oriented motor-sequences and habit learning (Graybiel, 1995; Barnes et al., 2005), whereas the NAc has been shown to be more involved in working-memory and reward-motivated behaviors and has been implicated in the development of addiction (Lovinger et al., 2003; Lewis et al., 2004). In the rodent brain, the striatum consists mainly of the caudate nucleus, whereas in the primate brain the striatum has differentiated into nucleus caudatus and putamen. Dorsal and ventral striatum likely have the same function within the basal ganglia-circuitry which is to translate cortical input into inhibition of downstream structures (Graybiel, 1998;

Lovinger et al., 2003).

Striatal input projections

The main inputs into the striatum arise from cortex, thalamus and midbrain (Fig. 1). In addition to this, smaller projection systems provide inputs into the striatum, these include; serotonergic input from the dorsal raphe nucleus and noradrenergic input from locus coeruleus.

Cortical input into the striatum originates from most cortical areas; motor, premotor and prefrontal regions as well as limbic cortical areas. The topographic organization of these cortical areas is largely maintained in their projections to the striatum and all cortico-striatal neurons are pyramidal neurons that use glutamate as a neurotransmitter (see Fig. 4). Three different corticostriatal cell subtypes have definitely been identified, these include: neurons of the pyramidal tract; bilaterally projecting corticocortical corticostriatal neurons; and neurons that are actually corticothalamic in nature but have a collateral projection to the striatum.

Thalamic input to the striatum was long believed to consist of a single topographically organized projection, but was later established to be heterogeneous in a similar way as

(10)

the corticostriatal projection pathways. In fact, two independent thalamostriatal projections were identified; one that originates from the parafascicular/centromedian nuclei and a separate one that arises from rostral parts of the complex. These projections are excitatory and use glutamate as a neurotransmitter.

Distinct groups of neurons located in the midbrain give rise to a dense efferent projection to striatum, limbic cortex and associated subcortical structures. These neurons use the catecholamine dopamine as a neurotransmitter and their cell-bodies are located primarily in the substantia nigra pars compacta (SNc) and ventral tegmental area (VTA).

Cellular composition of the striatum

Medium spiny neurons (MSNs) or spiny projection neurons are GABAergic neurons that constitute 90-95% of the cell-population of the striatum. MSNs have a cell body of approximately 20-25 µm in diameter from which extend 7-10 moderately branched dendrites that are densely laden with spines (Fig. 2). The distribution of the dendrites is not always uniform and may be limited by compartmental boundaries within the striatum as specified in a later paragraph. MSNs project their axons to the two major

Figure 1 – Sagittal section of rat brain showing striatal circuitry.

Green lines represent glutamatergic pathways, red lines represent GABAergic pathways and blue lines are dopaminergic projections. GABAergic (red) medium spiny neurons in the striatum (Nst) project to globus pallidus (GP) and substantia nigra (SN). Abbreviations – STR: striatum, CTX: cortex, Th: thalamus, STN: subthalamic nucleus, SN: substantia nigra, GP: globus pallidus. (Adapted from Greengard P et al. (1999) Neuron, Vol.23; 435- 447)

(11)

target structures of the striatum; the globus pallidus (GP) and the substantia nigra pars reticulata (SNr) (Fig. 1). Often, MSNs also sprout a local axon collateral that remains within the striatum but may extend over a very large area within this structure (Kawaguchi et al., 1990).

Three main classes of striatal interneurons have been identified based on their physiological, morphological and histochemical properties. The first type is a fast-spiking cell (FS-cell) that receives direct cortical input and is GABAergic in nature. This type of interneuron is thought to mediate feed-forward GABA- inhibition via the parvalbumin- positive terminals that make symmetrical synapses on somata and dendrites of neostriatal cells, probably including MSNs (Kita et al., 1990).

The second type of interneuron, the persistent and low-threshold spike cell (PLTS-cell) exhibits unique firing properties (Kawaguchi, 1993) and has NADPH diaphorase immunoreactivity. As NADPH diaphorase and NO-synthase are identical (Dawson et al., 1991), these neurons are considered to release NO within the striatum. They are positive for neuropeptide Y and are thought to be GABAergic. Like FS-cells, these cells regulate MSNs from the cortex through feed-forward inhibition. Thus, cortical information may affect MSNs through two types of interneurons.

The third type, the long-lasting afterhyperpolarization cell (LA-cell) is characterized by large afterhyperpolarizations which causes it to fire single spikes. They are cholinergic as they display ChAT-immunoreactivity, and they mainly receive input from the thalamus. Their terminals end on dendritic shafts and perikarya of MSNs in symmetrical synapses, and they have been proposed to stabilize the state of MSNs, whether it is depolarized or hyperpolarized (Akins et al., 1990).

Figure 2 – Cell body of a GABAergic, striatal medium spiny neuron (MSN) displaying a dendrite laden with spines and an output-axon.

Glutamatergic inputs from cortex synapse on the heads of spines and dopaminergic inputs from midbrain on the necks of spines and dendrites (see example in box). Interneuronal input can both be GABAergic and cholinergic. (Adapted from Arbuthnott GW et al. (2000) J. Anat. 196; 587-596)

(12)

Striatal output projections

Two parallel pathways run from the striatum to two output nuclei; globus pallidus (GP) and substantia nigra pars reticulata (SNr) (Fig. 1, 3). The first pathway, termed the indirect pathway, sends inhibitory, GABAergic projections to GP and from there to the subthalamic nucleus (STN). In the STN, the output signal is transformed into excitatory, glutamatergic projections that reach SNr. This projection from STN to SNr is the only excitatory projection in the basal ganglia. Striatal MSNs project directly to SNr in the second pathway. This inhibitory, GABAergic pathway is therefore called the direct pathway.

Immunohistochemical studies suggest that striatopallidal neurons (indirect pathway) express enkephalin (ENK), whereas striatonigral neurons (direct pathway) express substance P (SP) and dynorphin (DYN) (Gerfen, 1992). In addition, the two striatal output pathways are affected differently by the dopaminergic input from substantia nigra pars compacta (SNc). Striatonigral neurons containing SP and DYN also express the dopamine D1-receptor that facilitates transmission. The inhibitory dopamine D2- receptor is expressed only by striatopallidal neurons that co-express ENK (Gerfen et al., 1990; Le Moine et al., 1990). In spite of the differences in dopamine receptor expression, dopaminergic inputs into these pathways lead to the same effect in reducing inhibition in thalamocortical neurons. This reduced inhibition will facilitate movements

Figure 3 – Striatal output projections.

Cortex and thalamus provide excitatory input to the striatum. Striatal neurons that contain enkephalin (ENK) and the dopamine D2-receptor (D2) provide inhibitory input to globus pallidus (GP). GP sends inhibitory projections to subthalamic nucleus (STN) which in turn provides excitatory input to substantia nigra pars reticulata (SNr). Striatal neurons expressing the dopamine D1-receptor (D1), dynorphin (DYN) and substance P (SP) directly provide SNr and entopeduncular nucleus (EP) with inhibitory input.

GABAergic neurons from SN and EP provide inhibition to thalamus, superior colliculus and pedunculopontine nucleus (PPN). (Adapted from Gerfen et al. (1992))

(13)

initiated by the cortex. Thus, dopaminergic input into the striatum is essential for initiation of movement.

Patch-matrix compartments

The division of striatal neurons into two clearly identifiable output pathways that regulate movement-initiation via disinhibition of the thalamus gives a deceptive impression of the complexity of the regulatory mechanisms within the striatum itself.

The considerable heterogeneity in the striatum is organized in a mosaic structure composed of two interdigitating compartments differing in cytochemical make-up and afferent and efferent targets. MSNs in the larger of these compartments, the matrix compartment, express calbindin and a plexus of somatostatin-immunoreactive fibers.

These neurons preferentially receive inputs from sensory and motor cortices. The MSNs located in the smaller compartment, called patch or striosome, express µ-opiate receptors, display acetylcholinesterase labeling and receive input from a more restricted area of the cortex (Donoghue and Herkenham, 1986; Kawaguchi et al., 1990; Gerfen, 1992).

Retrograde axonal tracing studies show that both patch and matrix neurons project to the substantia nigra, but that patch neurons provide inputs to dopaminergic cells in both SNc and SNr whereas matrix neurons provide inputs to GABAergic neurons in SNr.

Thus, the striatonigral pathway is subdivided in a patch-matrix manner with patch- neurons innervating the dopaminergic parts of SN and matrix-neurons innervating the non-dopaminergic parts of SNr. Multiple studies have shown that dendrites of subsequent patch and matrix neurons do not extend beyond their separate compartments, suggesting that inputs are confined to the separate compartments as well (Bolam et al., 1988). Indeed, dopaminergic inputs from midbrain VTA, SN and retrorubral area provide inputs into the striatum that specifically target either the patch- or the matrix-compartments (Jimenez-Castellanos and Graybiel, 1987).

Segregation between dorsal and ventral striatum

The cortical, thalamic and dopaminergic inputs into the striatum do not segregate along well-defined lines. The boundary between the NAc (ventral striatum) and the caudate putamen (dorsal striatum) is widely used as a demarcation line, but a clear distinction between these two striatal subregions has not been identified (Groenewegen et al., 1999; Voorn et al., 2004). In line with the dorsal-ventral striatal division, appetitive

(14)

behavior and reinforcement are generally agreed to be ventral striatal functions (Cardinal et al., 2002; Kelley, 2004b). However, some effects of psychostimulants on conditioned reinforcement (Baker et al., 1998) and feeding behavior (Kelley, 2004a) can be ascribed to parts of the dorsal striatum, slightly blurring the dorsal-ventral division. Similar functional overlap can be seen in cognitive functions (Setlow, 1997;

Devan and White, 1999).

No clear boundary between dorsal and ventral striatum can be established based on cytoarchitecture or chemoarchitecture (Prensa et al., 2003). Instead, a dorsolateral to ventromedial graded density in MSN’s has been observed that resembles both the zonal organization of several neurochemical gradients and the pattern of corticostriatal inputs (Karachi et al., 2002; Haber, 2003). This dorsolateral to ventromedial distinction is also in line with regional differentiation of behavioral functions (Voorn et al., 2004).

Together, these findings suggest that a dorsolateral to ventromedial functional striatal organization would provide a better framework to define striatal boundaries than would the classic dorsal-ventral divide (Fig. 4).

This divide fits well with the already established distinction of the functionally related core and shell regions of the nucleus accumbens. These two regions would represent the ventral-most sector in the dorsolateral to ventromedial functional organization with

Figure 4 – Cortical and thalamic inputs to the striatum distribute in dorsomedial-to-ventrolateral zones (grayscales).

This distribution is illustrated by showing the topographical arrangement of afferents originating in the frontal cortex, midline and intralaminar thalamic nuclei, basal amygdaloid complex and hippocampal formation. Afferent projections from these regions converge in longitudinal striatal zones with a dorsomedial-to-ventrolateral orientation. The traditional distinction between dorsal and ventral striatal areas and, more specifically, the core and shell region of the nucleus accumbens are also outlined. Abbreviations of interest – ac, anterior commissure; SMC, sensorimotor cortex; PFC, prefrontal cortex (reproduced with permission; Voorn et al. 2004).

(15)

the ventral and medial extreme being the caudomedial shell. The shell reaches areas that are unique for striatal output such as preoptic, hypothalamic and mesencephalic areas associated with locomotor functions. So, although dividing the striatum into dorsal and ventral extremes has greatly contributed to gaining understanding of striatal function, it might be more appropriate to view striatal function from a mediolateral perspective (Voorn et al., 2004).

II. SYNAPTIC TRANSMISSION AT THE CORTICOSTRIATAL SYNAPSE Glutamatergic synaptic transmission

The major force driving synaptic transmission in the striatum is the excitatory input coming from cortex (corticostriatal pathway). These projections use the excitatory amino acid glutamate as a neurotransmitter. Glutamate is a nonessential amino acid that does not cross the blood-brain barrier and is therefore locally synthesized in neurons from precursors. The most abundant precursor for the synthesis of glutamate is glutamine, which is released into the extracellular space by glial cells. Glutamine is taken up by nerve terminals via excitatory amino acid transporters (EAATs) and converted to glutamate by the mitochondrial enzyme glutaminase in the cytoplasm.

After synthesis, glutamate is packaged into synaptic vesicles by vesicular glutamate transporters (VGLUT). Once glutamate is released into the synaptic cleft, it is taken up by glial cells via EAATs and broken down to glutamine by the enzyme glutamine synthetase. Glutamine is then released back into the extracellular space and can be reused by nerve terminals to synthesize glutamate. This cooperation between nerve terminals and glial cells to maintain a sufficient store of neurotransmitter is called the glutamate-glutamine cycle (Fig. 5).

Glutamate activates postsynaptic glutamate receptors. In the striatum, both ionotropic and metabotropic glutamate receptors are present. Ionotropic glutamate receptors are ligand-gated cation channels that are subdivided into three subtypes; NMDA (N- methyl-D-aspartate), AMPA (amino-3-hydroxy-5-methyl-4-isoazole propionic acid) and kainate (KA) receptors (Herrling et al., 1983; Tarazi and Baldessarini, 1999).

Evidence indicates that all three subtypes are expressed in the striatum. For instance, electrical stimulation of cortical inputs evokes excitatory postsynaptic potentials (EPSPs) in striatal MSNs which are largely mediated by AMPA receptors (Calabresi et al., 1990). In addition to this, evidence indicates a significant role for the NMDA- receptor in MSNs. MSNs are susceptible to toxic doses of NMDA or over-activation of

(16)

NMDA-receptors which may result in excitotoxicity and neuronal cell-death (Gerfen, 1992).

Metabotropic glutamate receptors (mGluRs) are coupled to various intracellular signal transduction processes. To date, eight mGluR subtypes have been cloned from the mammalian brain. These subtypes are classified into three major groups based on sequence homologies, coupling to second messenger systems and pharmacological profiles. Group I mGluRs, which include mGluR1 and mGluR5, couple primarily to Gq and increase phosphoinositide hydrolysis. Groups II (mGluR2 and mGluR3) and III (mGluR 4, 6, 7 and 8) couple to Gi/Go and inhibition of adenylyl cyclase (Conn and Pin, 1997). The Group I mGluRs, mGluR1 and 5, have been identified in striatal MSNs (Kerner et al., 1997). Activation of Group I mGluRs was shown to potentiate NMDA receptor currents in striatal neurons (Pisani et al., 1997; Rouse et al., 2000).

The NMDA-receptor

NMDA receptors are glutamate-activated cation channels that are characterized by a high Ca²⁺/Na⁺ permeability ratio. The pore of the channel is blocked by Mg²⁺ in a voltage-dependent manner and activation of the receptor requires glycine as a co agonist of glutamate (Dingledine et al., 1999). The NMDA-receptor mediates synaptic transmission and neural plasticity at many sites in the mammalian central nervous system. It has slow activation and deactivation kinetics and can contribute to

Figure 5 – Neurotransmission at the glutamatergic synapse.

In the presynaptic terminal glutamine is converted into glutamate (Glu) which is transported into synaptic vesicles. Influx of Ca²⁺triggers release of glutamate into the synaptic cleft, where it activates postsynaptic receptors; NMDA, AMPA, kainate and mGluR. Glutamate is cleared from the synaptic cleft by glutamate transporters that transport glutamate back into the terminal or into adjacent glial cells. In glial cells glutamate is converted into glutamine to fulfill the glutamate-glutamine cycle.

(17)

epileptiform activity and excitotoxicity leading to neuronal cell death in certain experimental and pathological conditions (Zeron et al., 2002; Jarabek et al., 2004).

NMDA receptors consist of an NR1 subunit and any of four NR2 subunits (NR2A-D) (Fig. 6). The NR1 subunit is expressed throughout the brain, whereas the expression of NR2 subunits is spatially and temporally regulated. Although all striatal MSNs express NR2A and NR2B subunits (Standaert et al., 1999), the relative expression of these subunits varies in different areas of the striatum. For instance, a distinct lateral to medial gradient of the NR2A mRNA distribution can be observed. In addition, NR2B was shown to be more abundantly expressed than NR2A (Buller et al., 1994). NMDA receptors are known to be located both intra- and extrasynaptically, fulfilling different roles with respect to development, synaptic plasticity and cell survival (Hardingham et al., 2002). NMDA-receptors with distinct pharmacologies appear to be involved in the release of the striatal neurotransmitters acetylcholine, dopamine and GABA (Nicolas et al., 1994; Nankai et al., 1995). Furthermore, striatal infusion of antisense oligonucleotide probes directed against NR2A and NR2B subunits produced differential effects in behavioral paradigms (Standaert et al., 1996).

On a pathophysiological level, NMDA receptor subunits are thought to be promising therapeutic targets for the development of new medication for treatment of several illnesses and disorders. Dysfunction of glutamatergic transmission in the striatum is believed to underlie pathologies such as Huntington’s disease, Parkinson’s disease and schizophrenia. For instance, animal models of Huntington’s disease demonstrate an increased vulnerability to NMDA-mediated cell death (Zeron et al., 2002). In animal models, NMDA receptor antagonists have proven to be effective antiparkinsonian

Figure 6 – Schematic presentation of the NMDA receptor.

The NR1 and NR2 subunits demonstrate N-terminal binding sites for glycine/D-serine and glutamate agonists and antagonists. Pore-opening requires the release of the Mg²⁺-ion by a depolarization of the membrane.

NMDA-receptors are often fixed in the post-synaptic membrane by PSD- scaffolding proteins. (Adapted from Kristiansen et al. 2007)

(18)

agents that have the potential to reduce application of other, more complicated types of therapeutics (Hallett and Standaert, 2004; Kristiansen et al., 2007).

Electrophysiology of striatal neurons

Intracellular recording studies have shown that the membrane potential of MSNs shifts periodically between a hyperpolarized ‘down’-state and a depolarized ‘up’-state (Calabresi et al., 1990). In the ‘down’-state, MSNs typically display a membrane potential of around -60 to -90 mV. In this stage MSNs are at resting membrane potential. When depolarized, MSNs display a membrane potential of approximately -60 to -40 mV. The ‘up’-state, or depolarized state, is generated by simultaneous activation of a large number of corticostriatal and thalamostriatal fibers (Wilson, 1993). During the ‘up’-state, the membrane potential reaches a spike threshold which triggers a train of spikes in the MSN. Thus, the synaptic inputs to the MSNs and their membrane- properties are major determining factors of their firing-potential (Calabresi et al., 1987).

In vivo intracellular recordings of MSNs have demonstrated that MSNs fire infrequently. Periods of membrane hyperpolarization and electrical silence are briefly interrupted by periods of sustained depolarization driven by cortical inputs (Wilson, 1993).

MSNs do not fire until a cortex command arrives. When they do fire, an inhibitory intrinsic and reciprocal network controls the propagation of excitation (Flores- Hernandez et al., 1994). This network may consist of several components.

Cholinergic interneurons form synapses on the dendrites of MSNs and intracellular recordings suggest that cholinergic synapses elicit both nicotinic and muscarinic responses in MSNs. In addition to this, GABAergic interneurons were shown to participate in a feedforward inhibition mechanism by forming synapses on the soma and proximal dendrites of MSNs (Kita et al.,

Figure 7 – Dopaminergic and glutamatergic synaptic transmission in striatal neurons. Dopaminergic terminals synapse on necks of spines and are therefore in an excellent position to modulate intracellular signaling pathways that may affect glutamatergic transmission.

(19)

1990). Intrastriatal stimulation was shown to evoke GABAergic inhibitory postsynaptic potentials (IPSPs) in MSNs that are mediated by GABAA-receptors (Kita, 1996).

Another important modulator of striatal MSNs is dopaminergic input from the midbrain. Dopaminergic terminals typically synapse on the necks of dendritic spines of MSNs. Dopamine receptors are known to regulate intracellular cAMP levels and all three subtypes of ionotropic glutamate receptors have been shown to be functionally modulated by activation of cAMP-dependent protein kinase (PKA) (Greengard et al., 1991). Anatomic studies have localized both ionotropic glutamate receptors and dopamine receptors to dendrites of MSNs, raising the possibility that dopamine receptor activation may regulate glutamate receptor function and thereby modulate the postsynaptic responsiveness of MSNs (Hersch et al., 1995).

Synaptic plasticity

Changes in the efficacy of synaptic transmission at synapses can contribute to storage of information within neural circuits. These changes in synaptic strength can occur both on a short-term and long-term basis depending on synaptic activity and the type of synapse (Bliss and Collingridge, 1993). Experimentally, it is possible to induce prolonged (>1 hour) increases and decreases in synaptic strength which have been termed long-term potentiation (LTP) and long-term depression (LTD), respectively.

The most extensively studied type of synaptic plasticity is NMDA-receptor dependent LTP and LTD in the CA1 region of the hippocampus (Bear and Malenka, 1994).

NMDA receptors have been shown to play an essential role in inducing long-term changes in synaptic strength (Malenka and Bear, 2004) and the subunit-composition of NMDA-receptors can be of importance for the induction of synaptic plasticity (Massey et al., 2004). Differential incorporation of NR2A or NR2B subunits within NMDA- receptor complex is thought to have profound implications for the generation of synaptic plasticity due to different interactions with proteins involved in intracellular cascades. The subunit-composition of NMDA receptors is under developmental regulation, and so is their consequential role in synaptic plasticity (Schramm et al., 2002). In fact, a developmental switch where NR2B predominates in early developmental stages and is gradually replaced by NR2A is associated with a decrease in the potential for generation of LTP due to the dramatically lower affinity for CAMKII of NR2A compared to NR2B (Barria and Malinow, 2005; Bayer et al., 2006).

(20)

The corticostriatal pathway was shown to support activity-dependent changes in synaptic plasticity (Partridge et al., 2000; Spencer and Murphy, 2000). Early studies on plasticity in striatal slices were primarily focused on LTD at corticostriatal synapses due to the relative ease with which this form of plasticity could be induced within the slice-preparation (Calabresi et al., 1992b). The induction of LTP in the dorsal striatum was generally thought to be difficult to achieve in slice-preparations until recent experiments revealed robust LTP in brain-slices using different recording techniques (Calabresi et al., 1992a). LTP-induction also seems to be dependent on the striatal subregion (Liu et al., 2004; Massey et al., 2004).

III. MODULATORS OF STRIATAL SYNAPTIC TRANSMISSION Adenosine

The purine and neuromodulator adenosine is not considered to be a classical neurotransmitter as it is not stored in synaptic vesicles or released in a Ca²⁺-dependent manner. Instead, it is generated from ATP by extracellular enzymes. Adenosine is thought to be involved in the regulation of important central mechanisms such as cognition, arousal, aggression and anxiety (Lang et al., 2003). In several brain regions, including brainstem, hippocampus and cortex, adenosine has been shown to be inhibitory (Greene and Haas, 1991). This inhibitory action was also demonstrated in the corticostriatal pathway (Malenka and Kocsis, 1988). In the striatum, endogenous adenosine levels are thought to be correlated with motor-activity (Huston et al., 1996).

Adenosine can influence physiological processes involving NMDA receptors, and shows interactions with dopamine receptors in the basal ganglia (Manzoni et al., 1994;

Fuxe et al., 1998). Adenosine has also been suggested to play a significant role in synaptic plasticity at specific synapses in the NAc and the striatum (Flagmeyer et al., 1997; d'Alcantara et al., 2001).

The adenosine A1-receptor

Transmission via purinergic synapses is thought to be widespread in the mammalian brain which is exemplified by a wide distribution of purinergic receptors, both ionotropic and metabotropic. The adenosine A1 receptor is one of four known subtypes of adenosine receptors: A1, A2A, A2B and A3. The A1 receptor is a G-protein coupled receptor that is linked to inhibition of transmitter release. Neuroprotection through the inhibition of glutamate-release is also ascribed to activation of this receptor (Rudolphi

(21)

et al., 1992). In the striatum, endogenous adenosine acting at adenosine A1-receptors may have neuroprotective effects during traumatic or metabolic stress where NMDA- receptors are involved (Calabresi et al., 1997; Centonze et al., 2001). In general, the A1 receptor is characterized as a homeostatic receptor with protective functions in many tissues.

The expression-levels of A1-receptors in striatum were shown to be intermediate to high whereas in NAc low levels were observed (Fastbom et al., 1987). In the striatum, adenosine A1 receptors are shown to be colocalized with D1 receptors in MSNs that project to substantia nigra (SNr) and entopeduncular nucleus (EP). In this pathway, adenosine A1 receptors were found to strongly antagonize dopamine receptor D1- mediated facilitation of GABA-release into the EP suggesting the existence of an antagonistic A1/D1 interaction in the direct pathway. A1 receptors might contribute to this effect by inhibiting glutamate release (Ferre et al., 1996). A1 receptors are also present in striatopallidal MSNs as well as in glutamatergic corticostriatal neurons (Ferre et al., 1997; Ferre et al., 1999).

Dopamine

Dopamine (DA) is one of the three well-established catecholamine neurotransmitters that are derived from the amino acid tyrosine. DA-synthesis occurs in the cytoplasm.

Tyrosine is converted into DOPA by the aid of tyrosine hydroxylase, an enzyme essential for DA-synthesis. DOPA is then converted into DA by the enzyme DOPA decarboxylase. Upon synthesis, DA is transported into vesicles by a vesicular monoamine transporter (VMAT). After release into the synaptic cleft a Na⁺-dependent dopamine transporter (DAT) ensures DA reuptake into nerve terminals and surrounding glial cells.

DA was first described in the central nervous system in 1964 (Dahlström and Fuxe, 1964). DA-ergic cell-bodies are shown to be primarily localized within the substantia nigra pars compacta (SNc) and ventral tegmental area (VTA). Their efferent projections terminate both in the dorsal striatum and NAc, as well as in limbic cortex and associated subcortical structures (Nauta et al., 1978; Voorn et al., 1986; Descarries et al., 1996). DA-ergic inputs importantly modulate synaptic transmission in the striatum.

Postsynaptic interactions between DA and glutamate in the NAc are shown to be critical for motivation, reward and other behavioral functions that are disrupted after chronic exposure to drugs of abuse (Everitt and Wolf, 2002). Ultrastructural studies

(22)

have shown that DA-ergic terminals form both symmetrical and asymmetrical contacts on postsynaptic structures (Sesack and Pickel, 1992) and are therefore in a position to modulate synaptic transmission by releasing DA in the vicinity of glutamatergic synapses or by causing more diffuse increases in extracellular DA levels (Grace, 1991;

Garris et al., 1994; Descarries et al., 1996).

Evidence has shown that striatal DA-transmission occurs in two distinct temporal modes; tonic and phasic. Phasic DA-transmission is a high-amplitude transient signal that consists of DA released by burst firing of DA-terminals. The transient nature of the signal is ensured via reuptake mechanisms mediated by DAT that eliminate DA from the synaptic cleft. This type of activity is rapid and behaviorally relevant. In contrast, tonic DA-transmission represents a pool of DA that is present at steady-state concentrations in the extra-cellular space. By activating autoreceptors located extrasynaptically on the DA-terminal, tonic DA-levels act to downregulate the responsivity of the DA-system to burst-firing generated during behavioral activation (Grace, 1991; Garris et al., 1994).

Dopamine-receptors

Originally, two types of DA-receptors (D1 and D2) were identified on pharmacological and biochemical grounds, based on their ability to activate (D1) and inhibit (D2) adenylate cyclase. Gene cloning, however, revealed five subgroups (D1-D5) where D1

and D₅ belong to the original D1-family and D₂, D₃ and D₄ fall into the initial D2- group. All subtypes are G-protein coupled transmembrane receptors and their transduction mechanisms are linked to adenylate cyclase and phospholipid hydrolysis.

Dopamine acts both pre- and postsynaptically. The impact of DA-receptor activation on MSN activity in the intact striatum is dependent on multiple factors, including the mode of DA-transmission (tonic or phasic), the relative contribution of D1 and D2 receptors and the striatal subregion involved. As was mentioned in previous chapters, dopaminergic input into the striatum is thought to modulate the response of MSNs to excitatory input by influencing voltage-gated ion channels, as well as ligand-gated ion channels such as the NMDA-receptor (Hernandez-Lopez et al., 1997).

D1-receptor mediated enhancement of NMDA-receptor responses was first described in the striatum about a decade ago and later confirmed in the nucleus accumbens (Cepeda and Levine, 1998; Chergui and Lacey, 1999). Different types of NMDA- receptors were shown to be involved in the D1-receptor mediated increase in

(23)

immediate early gene expression in striatonigral neurons (Keefe and Gerfen, 1996).

The interaction between NMDA-receptors and dopaminergic inputs also affects striatal synaptic plasticity as the activation of D1-receptors was shown to be required for the induction of NMDA-dependent LTP (Calabresi et al., 2000; Kerr and Wickens, 2001).

DA also modulates MSNs via inhibitory pathways which involves D2 receptor- dependent mechanisms. Studies in striatal brain slices revealed that DA or D2-agonists applied in the bath decreased the amplitude of evoked corticostriatal EPSPs and locally evoked EPSCs (Levine et al., 1996; Umemiya and Raymond, 1997). This D2 receptor- mediated inhibition of corticostriatal inputs is likely to be, at least partially, mediated by presynaptic receptors (O'Donnell and Grace, 1994; Hsu et al., 1995). In addition, MSNs from D2 receptor-deficient mice exhibited increased spontaneous synaptic activity and large amplitude depolarizations that were not observed in wild-type mice (Cepeda et al., 2001). Together, these studies demonstrate that it is likely that D2- receptors play a significant role in inhibiting corticostriatal signaling in striatal MSNs.

Estrogen

Traditionally, a distinction is made between hormones (chemicals that are released from glands into the blood-stream), and neurotransmitters (chemicals that are synthesized by neurons, stored in vesicles and exert their actions on post-synaptic receptors). This distinction, however, was challenged in the 1950s by the discovery of neurohypophyseal hormones such as oxytocin and vasopressin that are synthesized by neurons, but are released into the blood-stream (Harris, 1951). From that moment onwards, the concept of a neurotransmitter was liberalized which allowed for a more integrative perspective on the interaction between neuronal and endocrinological compartments. Recent work on brain steroids, particularly locally synthesized estrogen (E2), reveals that E2 in the brain displays many features that are traditionally associated with neurotransmitters. Indeed, estrogens are synthesized in the brain by aromatization (aromatase) of androgens such as testosterone. Rapid changes in aromatase-activity can be induced by phosphorylation and changes in intracellular Ca²⁺-levels. E2 was found to be present in pre-synaptic boutons, displays rapid effects at a cellular level and is subject to rapid inactivation mechanisms (Balthazart et al., 2001).

(24)

Estrogen-receptors

Two nuclear estrogen-receptors, ERα and ERβ, modulate gene-transcription when activated by estrogen-agonists. These receptors are transcription factors that dimerize upon ligand-binding. This dimer then gathers other factors and forms a receptor complex that binds to hormone response elements of target genes. Many of the cellular effects of estrogens are mediated by these nuclear or ‘genomic’ actions (Sanchez et al., 2002).

Accumulating evidence indicates ‘non-genomic’ membrane-effects of estrogens in the brain that might either be mediated by membrane-bound ERα and ERβ (Toran- Allerand, 2004) or by a novel, plasma membrane-associated,putative ER with unique properties (Toran-Allerand et al., 2002). In fact, nuclear and membrane ERs can be derived from a single gene transcript and display near-identical affinities for E2 (Razandi et al., 1999). Membrane-bound ERs can trigger activation of second messenger systems such as cAMP and IP3 production and CREB activation (Levin, 2002) as well as activation of MAP kinase (Wade et al., 2001).

The expression of mRNA for ERα and ERβ has been reported in the striatum (Kuppers and Beyer, 1999), and striatal expression of these two ERs is thought to be low to moderate (Shughrue et al., 1997; Zhang et al., 2002; Morissette et al., 2008). Several studies have demonstrated rapid effects of E2 in the striatum. For instance, in vitro administration of E2 rapidly stimulates striatal DA-release (Becker, 1990), and pre- treatment with E2 was shown to increase the firing rate of MSNs in response to DA (Arnauld et al., 1981). Indeed, such effects of E2 in the striatum have been associated with the activation of second messenger pathways via interaction of E2 with membrane-associated ERs (Mermelstein et al., 1996).

GABA

The inhibitory neurotransmitter γ-aminobutyric acid (GABA) was identified in brain tissue during the 1950s (Roberts and Frankel, 1950; Udenfriend, 1950). It is now known that as many as a third of the synapses in the brain use GABA as their neurotransmitter. GABA is commonly found in local circuit interneurons as can be identified in the striatum. Importantly, striatal MSNs are GABAergic projection- neurons that exert an inhibitory influence on their target-structures.

The predominant precursor for GABA synthesis is glucose, but pyruvate and glutamine can also act as precursors. These precursors are metabolized into glutamate and

(25)

converted into GABA by the catalyzing enzyme glutamic acid decarboxylase (GAD).

This enzyme is almost exclusively localized in GABAergic neurons and is therefore commonly used as a specific marker for these neurons. Once GABA is synthesized, it is transported into synaptic vesicles. Removal upon release is carried out by high- affinity transporters for GABA (GATs) that transport GABA back into neurons or glial cells.

GABA-receptors

Three types of postsynaptic GABA-receptors (GABAA-C) have been characterized, two of which are ionotropic receptors (GABAA and C) and one of which is metabotropic (GABAB). Ionotropic GABA-receptors are permeable to Cl^- and are therefore mostly inhibitory, since in most neurons the reversal potential for Cl^- is more negative than the threshold for neuronal firing. These receptors are pentamers assembled from five types of subunits (α, β, γ, δ and ρ) giving rise to a diverse range of combinations. As a consequence, the function of ionotropic GABA-receptors varies widely among neuronal types.

GABAA-receptors are essential for the function of the basal ganglia providing fast synaptic inhibition between basal ganglia nuclei (Smith et al., 1998). Most of the GABA-effects in the striatum are mediated by this receptor. Tonic GABA-inhibition due to low concentrations of GABA diffusing from the synapse and activating extrasynaptic GABA_A-receptors also occurs in basal ganglia circuits (Goetz et al., 2007). In the striatum, GABAA-receptors mediate fast transmission at symmetrical synapses on dendrites and spines of MSNs and GABA interneurons. They are also present on GABAergic as well as non-GABAergic pre-synaptic terminals (Fujiyama et al., 2000).

Like ionotropic GABA-receptors, metabotropic GABA-receptors (GABA_B) are inhibitory and widely distributed in the brain. GABAB-mediated inhibition is mostly due to the activation of K⁺-channels, but GABAB-mediated inhibition can also be due to blockade of Ca⁺-channels leading to inhibition of transmitter-release if GABAB- receptors are located pre-synaptically. GABAB-receptors are heterodimers of two subunits GABA_B1 and GABA_B2 (Kaupmann et al., 1998). In the striatum, GABA_B- receptors are located both pre- and postsynaptically. Presynaptic GABAB-receptors can modulate input-projections from cortex and thalamus and are present on terminals of

(26)

MSNs and GABAergic interneurons. Furthermore, postsynaptic GABAB-receptors may affect excitability of striatal neurons (Lacey et al., 2005).

(27)

AIMS

The general aim of the studies brought together in this thesis was to identify the role of several neurotransmitters; glutamate, dopamine and GABA as well as the neuromodulator adenosine in synaptic transmission and plasticity in the striatum of the mouse. Specific focus was placed upon the NMDA-receptor; its activation, its subunits and its role in synaptic plasticity.

The specific aims of this thesis are:

• To investigate the effect of NMDA on synaptic transmission and long-term synaptic plasticity and the interaction between NMDA-receptors and the neuromodulator adenosine in the striatum

• To investigate if the NR2A and NR2B subunits of the NMDA-receptor contribute to NMDA-induced synaptic depression of glutamatergic transmission and evoked dopamine-release in the striatum

• To examine the contribution of NMDA-receptor subunits NR2A and NR2B to high frequency stimulation (HFS)-induced long term potentiation (LTP) in the nucleus accumbens (NAc)

• To identify the contribution of the neurotransmitters dopamine, glutamate and GABA to high frequency stimulation (HFS)-induced long term potentiation (LTP) in the nucleus accumbens (NAc)

• To identify gender-related differences in synaptic transmission in the striatum and NAc and to identify the acute effect of estrogen (E2)

(28)

MATERIALS AND METHODS

Animals and brain slice preparation

Experiments using mice were approved of by our local ethical committee (Stockholms norra djurförsöksetiska nämnd). Male and female C57Bl6 mice aged 4-7 weeks (females 5-7 weeks) were used in all experiments on wild-type mice. Knockout mice with a deletion in the A1 receptor gene were back-bred against C57Bl6 in the speed congenic protocol of Jackson laboratories and judged to be congenic by 140 gene markers (Johansson et al., 2001). A1 receptor knockout mice used in the experiments were 4-7 week old males.

Mice were anaesthetized by inhalation of fluothane and subsequently underwent cervical dislocation and decapitation. The brain was swiftly removed and coronal striatal slices (400 µm thick) were prepared in ice cold aCSF on a microslicer (Leica, VT1000S). In the experiments using decorticated striatal slices, the cortex was removed from each slice immediately after sectioning. The slices were then equilibrated during at least 1 hour in carbogen rich (95:5) artificial cerebrospinal fluid (aCSF) (NaCl 126 mM, KCl 2,5 mM, NaH2PO4 1,2 mM, MgCl2 1,3 mM, CaCl2 2,4 mM, glucose 10 mM and NaHCO3 26 mM, pH 7.4) at a constant temperature of 32°C. Using a glass Pasteurs’ pipette filled with oxygenated aCSF, the slices were transferred to a recording chamber (Warner Instruments, Hamden, CT) mounted on an upright microscope (Olympus, Solna, Sweden) and were continuously perfused with oxygenated aCSF.

Identification of the estrous cycle of female mice

Since the first characterization of the estrous cycle in rats by Long and Evans (Evans and Long, 1922) the evaluation of changes in epithelial cell structure (vaginal smear) in spontaneously ovulating laboratory animals has been used as a principal approach to measure reproductive cyclicity. In juvenile mice a fully cornified smear indicates the onset of the first ovulation (around 5 weeks). The following cycles will tend to average 5 days, although irregularities may occur due to environmental conditions. For instance, female mice housed in the same room as male mice exhibit more regular cycles than females housed in an all-female environment. In addition, continuous illumination can induce a persistent estrus (Goldman et al., 2007).

(29)

In a standard cycle, proestrus is identified by the presence of round, nucleated epithelial cells, which often have a granular appearance under the microscope (Fig. 8A).

Proestrus lasts for one day and is followed by vaginal estrus, which is routinely identifiable by large numbers of cornified (or keratinized) cells which have either a needle-like or a more rounded shape (Fig. 8B). The predominance of cornified cells can last one or two days, depending on the length of the cycle (4-5 days). Metestrus reflects the transitional period during the early part of diestrus, and its smear is characterized by a combination of leukocytes and cornified and rounded epithelial cells (Fig. 8C, D).

During diestrus, the smear can often be almost exclusively leukocytic (Fig. 8E), but may also show a few small clumps of nucleated epithelial cells that indicate initiation of proestrus the next day.

The appearance of different cell-types in the vaginal smear is a reflection of the hormonal changes that are taking place in the rodent. The two main hormones that regulate the cyclicity in the female reproductive system are estrogen and progesterone.

Overall, estrogen promotes proliferation and growth of endometrial cells and increases

Figure 8 – Stages of the mouse estrous-cycle identified by means of vaginal smear samples.

A – Pro-estrus; characterized by a large amount of nucleated cells (white arrows), B – Estrus; characterized by cornified cells (black arrows), C, D – Met-estrus; a combination of nucleated cells and leukocytes (C, black arrows) and cornified cells (D, white arrow), E – Di-estrus; predominating leukocytes (black arrows).

(30)

vascular permeability. Estrogen also stimulates the proliferation and differentiation (keratinization) of the squamous epithelium of the vagina (Li and Davis, 2007).

Progesterone exerts an opposite effect of estrogen and leads to a decrease in the number of keratinized cells (Goldman et al., 2007; Li and Davis, 2007). Different studies report slight differences as to the exact moment when the fluctuations in hormone- concentrations take place. However, it can generally be stated that the levels of estrogen are high during proestrus, low during estrus and metestrus and intermediate during the diestrus-phase of the cycle. The differences observed in vaginal cytology are a relevant reflection of the circulating estrogen-levels, and therefore also of the hormone-status in the brain.

We collected vaginal smear samples from the female mice used in the electrophysiology-experiments by means of a micro-spatula (Fisher Scientific). The samples were fixed in 100% ethanol and were stained according to the Nissl-staining protocol (Cresyl violet). The samples were then identified on a regular light-microscope (Motic) by at least two separate, unbiased researchers.

Electrophysiology

Experiments were performed in a submerged-type chamber perfused with carbogen rich (95:5) aCSF. Extracellular field potentials were recorded using a borosilicate glass micropipette (low resistance; G120F-3; Warner Instrument Corp.) filled with aCSF positioned on the slice surface. A concentric bipolar stimulating electrode (FHC, Bowdoinham, ME) was placed in close proximity to stimulate fibers. Signals were amplified 500 times via an Axopatch 200B amplifier (Axon Instruments, Foster City, CA), acquired at 10 kHz, filtered at 2 kHz and recorded on a Dell computer using acquisition and data analysis software from Axon Instruments (pClamp9).

To evoke synaptic responses single stimuli (0.1 ms duration) were applied every 15 s at an intensity of 50-70% of the maximal response as established by a stimulus/response curve, for each slice, at the beginning of a recording session. The maximal stimulation intensity was determined measuring the amplitude of the field potential evoked by gradually increasing stimulation intensities. Baseline responses were recorded for at least 15-20 minutes. If the fEPSP slope was observed to drift by more than 25% during baseline recording the pulse intensity was re-adjusted and another baseline was recorded. If the fEPSP slope continued to drift the slice was abandoned.

(31)

Amperometry

Amperometric detection of dopamine release was performed with methods described earlier (Schmitz et al., 2001; Chergui et al., 2004). A carbon fiber electrode (WPI, 10µm diameter) with an active part of 100 µm was positioned within the striatum in the brain slice. A constant voltage of +500 mV was applied to the carbon fiber via an Axopatch 200B amplifier (Axon Instruments) and currents were recorded with the same amplifier. A stimulating electrode (patch micropipette filled with aCSF) was placed on the slice surface in close proximity to the carbon fiber electrode. Stimulation consisted of a single pulse (0.1 ms, 8-14 µA) applied every minute, which evoked a response corresponding to oxidation of dopamine at the surface of the electrode. When the carbon fiber electrode was held at 0 mV, stimulation of the slice did not produce any current.

Drugs and chemicals

In the experiments performed in this thesis, acute bath-administration of different kinds of drugs is an important tool. Drugs were dissolved according to the solubility-protocol.

Stock solutions of most drugs were stored at -20°C and fresh solutions were prepared by diluting with aCSF before each experiment. Drug-application to the slices was controlled by switching a three-way tap. The perfusion flow rate was maintained at 1- 1.5 ml/min.

Drugs targeting the glutamatergic system

NMDA

(N-methyl-D-aspartate, Sigma)

prototypic NMDA-receptor agonist

(Study I & II) (Carter et al., 1989)

DL-APV

(DL-2-amino-5-phosphonopentanoic acid, Tocris)

unselective, potent NMDA antagonist

(Study III)

Ifenprodil*

(2-(4-benzylpiperidino)-1-(4-hydroxy-phenyl)-1- propanol (hemi) tartrate, Tocris & Sigma)

selective antagonist for NR1/NR2B-diheteromeric receptors

(Study II & III)

Ro 25-6981* selective antagonist for NR1/NR2B-diheteromeric

(32)

([R-(R*,S*)] - α - (4-hydroxyphenyl) - β - methyl - 4 - (phenylmethyl) - 1 - piperidinepropanol – hydrochloride, Sigma)

receptors, derivative of ifenprodil

(Study II & III)

NVP-AAM 077**

([(R)-(S)-1-(4-bromo - phenyl) – ethyl-amino - (2,3-dioxo-1,2,3,4-tetrahydro-quinoxalin-5-yl) - methyl] - phosphonic acid, Novartis Pharma)

specific antagonist of NR1/NR2A-diheteromeric receptors (Study II & III)

CNQX

(6-cyano-7-nitroquinoxaline-2,3-dione disodium salt, Sigma)

potent, competitive AMPA/kainate receptor antagonist (Study I)

MPEP hydrochloride

(2-methyl-6-(phenylethynyl) pyridine hydrochloride, Tocris)

potent and highly selective non-competitive antagonist of the mGlu5 receptor subtype

(Study IV)

LY 367385

((S)-(+)-a-amino-4-carboxy-2- methylbenzeneacetic acid, Tocris)

selective mGlu1a receptor antagonist

(Study IV)

DL-TBOA

(DL-threo-b-benzyloxyaspartic acid, Tocris)

potent, competitive, non-transportable blocker of excitatory amino acid transporters with high selectivity for EAATs

(Study IV)

* Ifenprodil and Ro 25-6981 are structurally closely related and are suggested to exert their antagonistic action on the LIVBP-like domain (leucine/isoleucine/valine-binding protein-like domain) of the NR2B subunit (Perin-Dureau et al., 2002; Malherbe et al., 2003). The specific affinity of these drugs in blocking NR2B (Ro 25-6981; IC50 ± 0.02 µM and Ifenprodil; IC50 ± 0.4 µM) exceeds that for NR2A (Ro 25-6981; IC50 ± 52 µM and Ifenprodil; IC50 ± 49 µM) (Fischer et al., 1997).

** NVP AAM077 was a gift from Dr. Y.P. Auberson, Novartis Pharma, Basel, Switzerland.

The affinity of NVP-AAM 077 for NR2A (IC50 = 14nM) clearly exceeds that for NR2B (IC50

= 1800nM) (Auberson et al., 2002).

(33)

Drugs targeting the dopaminergic system

SCH 23390

(R(+)-SCH-23390 hydrochloride, Sigma)

D₁ dopamine receptor antagonist

(Study IV)

Sulpiride

((S)-(−)-Sulpiride, Sigma)

D2 dopamine receptor antagonist

(Study IV)

Nomifensine

(1,2,3,4-tetrahydro-2-methyl-4-phenyl-8- isoquinolinamine maleate salt, Sigma)

selective dopamine uptake inhibitor, interacts with the dopamine transporter at a site different from that of cocaine

(Study IV) (Wieczorek and Kruk, 1994)

Drugs targeting purinergic neurotransmission

Adenosine

(adenosine, Calbiochem)

endogenous neurotransmitter, purine

(Study I) DPCPX

(8-cyclopentyl-1,3-dipropylxanthine, Tocris)

potent and selective A₁ adenosine receptor antagonist (Study I)

Drugs targeting the GABAergic system

Bicuculline

((-)-Bicuculline methiodide, Sigma)

GABAA receptor antagonist

(Study II & IV)

CGP 55845

((2S) – 3 - [[(1S) – 1 - ( 3,4 – dichlorophenyl ) ethyl ] amino – 2 – hydroxypropyl ] (phenylmethyl) phosphinic acid, Tocris)

potent, selective GABAB receptor antagonist

(Study IV)

(34)

RESULTS AND DISCUSSION

Effects of NMDA-receptor activation on glutamatergic synaptic transmission and evoked-dopamine release in the striatum

In order to examine glutamatergic

synaptic transmission in the striatum we performed recordings of extracellular field potentials in acute corticostriatal brain slices. In earlier

studies a brief electrical stimulation of corticostriatal slices was shown to elicit a characteristic biphasic response (Fig. 9) with two negative components (Calabresi et al., 1997). The first negative component, termed N1, is a non-synaptic compound action potential which is independent of glutamate release as it is unaffected by the AMPA- receptor antagonist CNQX (Study I). The second negative component, termed “field excitatory postsynaptic potential/population spike“ (fEPSP/PS), is mediated by endogenous glutamate released upon electrical stimulation of glutamate-containing fibers present in the slice. In the striatum, this component is mostly mediated by glutamate receptors of the AMPA type as blocking these receptors with CNQX resulted in a full blockade of the fEPSP/PS (Study I).

NMDA depresses striatal glutamatergic synaptic transmission

Apart from AMPA-receptors, glutamate also activates NMDA-receptors to mediate fast synaptic transmission. NMDA-receptors are known to play a role in synapse development and neurotoxicity and are critically involved in certain forms of synaptic plasticity (Malenka and Bear, 2004). In the hippocampus, exogenous application of NMDA was shown to cause an initial inhibition of glutamatergic synaptic transmission followed by LTD (Lee et al., 1998). The ability of NMDA to induce long-term changes in glutamatergic transmission in the striatum has not yet been extensively studied. We

Figure 9 - Stimulation of striatal glutamatergic fibers results in a biphasic response with two negative components; N1 and fEPSP/PS.

(35)

therefore wanted to examine the ability of NMDA to regulate glutamatergic synaptic transmission and to induce long-term changes in synaptic strength in the striatum.

As we set out to evaluate the effect of NMDA-receptor activation on glutamatergic synaptic transmission we applied NMDA in the perfusion solution at various concentrations while measuring fEPSP/PS amplitude. We observed a depression of the fEPSP/PS in response to NMDA-application which was concentration-dependent and culminated in a full blockade at 40µM which lasted for 1 hour. At 32°C, this depression of fEPSP/PS was not a form of LTD as the fEPSP/PS returned to baseline levels after 2 hours. However, at room temperature (RT) the depression was longer lasting and was therefore identified as a form of LTD (Study I). These results show that NMDA profoundly disrupts glutamatergic synaptic transmission in a concentration- and temperature-dependent manner. Indeed, previous studies have indicated that NMDA- application at lower concentrations than we used in our study is able to induce a chemical form of LTD in hippocampus and dentate gyrus (Lee et al., 1998; Rush et al., 2001). However, such LTD is not induced in the striatum at a physiologically relevant temperature but only at RT suggesting that different expression mechanisms are activated at different temperatures.

NMDA-induced depression is mediated by adenosine via A1-receptor activation Activation of NMDA-receptors has been shown to cause the release of adenosine in cortex and hippocampus (Craig and White, 1993; Manzoni et al., 1994). We now wanted to test the possibility that the observed actions of NMDA might be mediated at least in part by the release of this purine. Of the four adenosine-receptors identified (Fredholm et al., 2001), the A1-receptor is the most likely candidate to mediate such an effect as it was shown to be involved in NMDA-mediated downregulation of glutamatergic synaptic transmission in hippocampus (Dunwiddie and Fredholm, 1989).

It is also highly expressed in the cortex (Mahan et al., 1991) and might therefore mediate adenosine-effects at corticostriatal terminals.

We investigated whether the NMDA-induced decrease in fEPSP/PS was mediated by the adenosine A1-receptor by examining the effect of NMDA in the presence of the selective A₁-receptor antagonist DPCPX and by studying this effect in A₁-receptor knockout mice (Study I). We found that the A1-receptor antagonist DPCPX dramatically reduces NMDA-induced synaptic depression both at 32°C and RT.

(36)

Moreover, the LTD we observed when applying NMDA (40µM) at RT was abolished by this intervention.

However, the initial depression induced by NMDA at 40µM remained largely unaffected. It is possible that incomplete blockade of A1-receptors was responsible for this remaining inhibition and we therefore repeated the experiments in brain slices from A1-receptor knockout mice. We found that the effect of NMDA at 40µM was significantly reduced and that the initial inhibition that we saw in wild-type slices with DPCPX was also dramatically attenuated, although not completely abolished. This observation suggests that post-synaptic membrane depolarization might account for a small component of NMDA-induced short term depression.

Finally, we wanted to find out whether adenosine itself would be capable of producing a similar effect to that induced by NMDA. We applied adenosine in the perfusion- solution at different concentrations and found a depression of the amplitude of the fEPSP/PS both at 32°C and RT. Similar to the NMDA-effect, the depressant effect of adenosine was larger at RT than at 32°C. The depressant effect of adenosine was absent in slices from A1-receptor knockout animals at both temperatures examined. These results clearly demonstrate that NMDA-receptors and adenosine A1-receptors cooperate to downregulate glutamate release in the striatum (Study I).

NMDA-induced depression is mediated by NMDA-receptors containing the NR2A-subunit

The precise molecular composition of functional NMDA-receptors in the striatum is currently unknown. As we now identified a specific mechanism involving NMDA- receptors which leads to down-regulation of striatal glutamate release (Study I), we wanted to identify the subunit-composition of these NMDA-receptors. Several studies have identified the subunit composition of striatal NMDA-receptors using different biochemical approaches and have found that the NR2B-subunit is most abundantly expressed (Standaert et al., 1994; Dunah and Standaert, 2003). The NR2A-subunit is also expressed in the striatum, but at more moderate levels while NR2C and NR2D- subunits are virtually absent in this brain region (Standaert et al., 1994; Landwehrmeyer et al., 1995; Ghasemzadeh et al., 1996). We therefore wanted to determine whether NR2B and/or NR2A subunits contribute to NMDA-induced synaptic depression in the striatum.

Synaptic transmission and synaptic plasticity in the mouse striatum

From DEPARTMENT OF PHYSIOLOGY AND PHARMACOLOGY

Karolinska Institutet, Stockholm, Sweden